👩‍💻Chord Graph

HoloView를 이용하여 Chord Graph 그리기

https://towardsdatascience.com/plotting-chord-diagrams-in-python-72fd71b3eef0

import pandas as pd

df = pd.read_csv('../flights.csv')
df

/home/jungyeon/anaconda3/envs/imi/lib/python3.7/site-packages/IPython/core/interactiveshell.py:3020: DtypeWarning: Columns (7,8) have mixed types.Specify dtype option on import or set low_memory=False.
  interactivity=interactivity, compiler=compiler, result=result)

	YEAR	MONTH	DAY	DAY_OF_WEEK	AIRLINE	FLIGHT_NUMBER	TAIL_NUMBER	ORIGIN_AIRPORT	DESTINATION_AIRPORT	SCHEDULED_DEPARTURE	...	ARRIVAL_TIME	ARRIVAL_DELAY	DIVERTED	CANCELLED	CANCELLATION_REASON	AIR_SYSTEM_DELAY	SECURITY_DELAY	AIRLINE_DELAY	LATE_AIRCRAFT_DELAY	WEATHER_DELAY
0	2015	1	1	4	AS	98	N407AS	ANC	SEA	5	...	408.0	-22.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
1	2015	1	1	4	AA	2336	N3KUAA	LAX	PBI	10	...	741.0	-9.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
2	2015	1	1	4	US	840	N171US	SFO	CLT	20	...	811.0	5.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
3	2015	1	1	4	AA	258	N3HYAA	LAX	MIA	20	...	756.0	-9.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
4	2015	1	1	4	AS	135	N527AS	SEA	ANC	25	...	259.0	-21.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
5819074	2015	12	31	4	B6	688	N657JB	LAX	BOS	2359	...	753.0	-26.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
5819075	2015	12	31	4	B6	745	N828JB	JFK	PSE	2359	...	430.0	-16.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
5819076	2015	12	31	4	B6	1503	N913JB	JFK	SJU	2359	...	432.0	-8.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
5819077	2015	12	31	4	B6	333	N527JB	MCO	SJU	2359	...	330.0	-10.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN
5819078	2015	12	31	4	B6	839	N534JB	JFK	BQN	2359	...	442.0	2.0	0	0	NaN	NaN	NaN	NaN	NaN	NaN

5819079 rows × 31 columns

• ORIGIN_AIRPORT
• DESTINATION_AIRPORT

Finding the Relationships between Origin and Destination Airports

df_between_airports = df.groupby(by=["ORIGIN_AIRPORT", "DESTINATION_AIRPORT"]).count()
df_between_airports

		YEAR	MONTH	DAY	DAY_OF_WEEK	AIRLINE	FLIGHT_NUMBER	TAIL_NUMBER	SCHEDULED_DEPARTURE	DEPARTURE_TIME	DEPARTURE_DELAY	...	ARRIVAL_TIME	ARRIVAL_DELAY	DIVERTED	CANCELLED	CANCELLATION_REASON	AIR_SYSTEM_DELAY	SECURITY_DELAY	AIRLINE_DELAY	LATE_AIRCRAFT_DELAY	WEATHER_DELAY
ORIGIN_AIRPORT	DESTINATION_AIRPORT
10135	10397	83	83	83	83	83	83	83	83	83	83	...	83	83	83	83	0	13	13	13	13	13
	11433	71	71	71	71	71	71	71	71	71	71	...	71	71	71	71	0	15	15	15	15	15
	13930	72	72	72	72	72	72	72	72	72	72	...	72	72	72	72	0	12	12	12	12	12
10136	11298	189	189	189	189	189	189	189	189	182	182	...	182	181	189	189	7	21	21	21	21	21
10140	10397	86	86	86	86	86	86	86	86	86	86	...	86	86	86	86	0	6	6	6	6	6
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
XNA	SFO	52	52	52	52	52	52	52	52	51	51	...	51	51	52	52	1	14	14	14	14	14
	SLC	1	1	1	1	1	1	1	1	1	1	...	1	1	1	1	0	0	0	0	0	0
YAK	CDV	331	331	331	331	331	331	331	331	328	328	...	325	325	331	331	3	42	42	42	42	42
	JNU	331	331	331	331	331	331	331	331	329	329	...	325	325	331	331	2	29	29	29	29	29
YUM	PHX	1877	1877	1877	1877	1877	1877	1877	1877	1854	1854	...	1854	1854	1877	1877	23	200	200	200	200	200

12377 rows × 29 columns

df_between_airports = df_between_airports['YEAR'].rename('COUNT').reset_index() 
df_between_airports

	ORIGIN_AIRPORT	DESTINATION_AIRPORT	COUNT
0	10135	10397	83
1	10135	11433	71
2	10135	13930	72
3	10136	11298	189
4	10140	10397	86
...	...	...	...
12372	XNA	SFO	52
12373	XNA	SLC	1
12374	YAK	CDV	331
12375	YAK	JNU	331
12376	YUM	PHX	1877

12377 rows × 3 columns

df_between_airports = df_between_airports.query('ORIGIN_AIRPORT.str.len() <= 3 & DESTINATION_AIRPORT.str.len() <= 3')
df_between_airports

	ORIGIN_AIRPORT	DESTINATION_AIRPORT	COUNT
7684	ABE	ATL	898
7685	ABE	DTW	711
7686	ABE	ORD	665
7687	ABI	DFW	2329
7688	ABQ	ATL	801
...	...	...	...
12372	XNA	SFO	52
12373	XNA	SLC	1
12374	YAK	CDV	331
12375	YAK	JNU	331
12376	YUM	PHX	1877

4693 rows × 3 columns

df_between_airports = df_between_airports.sort_values(by="COUNT",ascending=False)
df_between_airports

	ORIGIN_AIRPORT	DESTINATION_AIRPORT	COUNT
11867	SFO	LAX	13744
10165	LAX	SFO	13457
9949	JFK	LAX	12016
10130	LAX	JFK	12015
10053	LAS	LAX	9715
...	...	...	...
11726	SBN	COS	1
10329	MCI	AUS	1
8139	BOI	EUG	1
9688	IAD	TTN	1
10227	LGA	EYW	1

4693 rows × 3 columns

top = 40
df_between_airports.head(top)['ORIGIN_AIRPORT'].unique()

array(['SFO', 'LAX', 'JFK', 'LAS', 'LGA', 'ORD', 'OGG', 'HNL', 'ATL',
       'MCO', 'DFW', 'SEA', 'BOS', 'DCA', 'FLL', 'PHX', 'DEN', 'TPA'],
      dtype=object)

df_between_airports.head(40)

	ORIGIN_AIRPORT	DESTINATION_AIRPORT	COUNT
11867	SFO	LAX	13744
10165	LAX	SFO	13457
9949	JFK	LAX	12016
10130	LAX	JFK	12015
10053	LAS	LAX	9715
10253	LGA	ORD	9639
10132	LAX	LAS	9594
11085	ORD	LGA	9575
11864	SFO	JFK	8440
9974	JFK	SFO	8437
10925	OGG	HNL	8313
9536	HNL	OGG	8282
10149	LAX	ORD	8256
7887	ATL	LGA	8234
10204	LGA	ATL	8215
7892	ATL	MCO	8202
10372	MCO	ATL	8202
11866	SFO	LAS	7995
11082	ORD	LAX	7941
10083	LAS	SFO	7870
8989	DFW	ORD	7870
11797	SEA	LAX	7765
10164	LAX	SEA	7732
8168	BOS	DCA	7687
8693	DCA	BOS	7686
11044	ORD	DFW	7677
7854	ATL	FLL	7419
9337	FLL	ATL	7411
11342	PHX	LAX	7380
11887	SFO	ORD	7380
11138	ORD	SFO	7378
10153	LAX	PHX	7330
8851	DEN	PHX	7211
11322	PHX	DEN	7204
10208	LGA	BOS	7100
8182	BOS	LGA	7096
12220	TPA	ATL	7083
7955	ATL	TPA	7076
8891	DFW	ATL	7060
7839	ATL	DFW	7056

Displaying the Chord Diagram

import holoviews as hv
hv.extension('bokeh')

/home/jungyeon/anaconda3/envs/imi/lib/python3.7/site-packages/requests/__init__.py:104: RequestsDependencyWarning: urllib3 (1.26.9) or chardet (5.0.0)/charset_normalizer (2.0.12) doesn't match a supported version!
  RequestsDependencyWarning)

HoloViews는 셀의 출력 결과를 표시하는 방법을 수정하기 위해 %%opts 셀 매직을 사용합니다. Chord 클래스를 사용하여 코드 다이어그램을 표시합니다.

코드 다이어그램에서 각 원(노드로 불리는)은 공항을 나타냅니다. 공항 간의 관계를 보려면, 원 위에 마우스를 올리면 됩니다.

%%opts Chord [height=500 width=500 title="Flights between airports" ]
chord = hv.Chord(df_between_airports.head(top))
chord

출력된 코드 다이어그램에서는 목적지 공항이 무엇인지 알아볼 수 없는 것 같습니다. 따라서, 출발지와 도착지 공항의 목록을 가져와 hv.Dataset 객체를 생성하는 데 사용하겠습니다.

# get the top count of flights between airports
df_between_airports = df_between_airports.head(top)

# find all the unique origin and destination airports
airports = list(set(df_between_airports["ORIGIN_AIRPORT"].unique().tolist() + 
                    df_between_airports["DESTINATION_AIRPORT"].unique().tolist()))

airports_dataset = hv.Dataset(pd.DataFrame(airports, columns=["Airport"]))

각 노드에서 공항 이름을 표시하려면, %%opts 셀 매직에서 labels 속성을 설정하고 airports_dataset 변수를 Chord 클래스 초기화자로 전달하면 됩니다.

%%opts Chord [height=500 width=500 title="Flights between airports" labels="Airport"]
chord = hv.Chord((df_between_airports, airports_dataset))
chord

참고할 점은 df_between_airports와 airports_dataset 변수가 튜플로 묶여 있다는 것입니다.

(df_between_airports, airports_dataset)

(      ORIGIN_AIRPORT DESTINATION_AIRPORT  COUNT
 11867            SFO                 LAX  13744
 10165            LAX                 SFO  13457
 9949             JFK                 LAX  12016
 10130            LAX                 JFK  12015
 10053            LAS                 LAX   9715
 10253            LGA                 ORD   9639
 10132            LAX                 LAS   9594
 11085            ORD                 LGA   9575
 11864            SFO                 JFK   8440
 9974             JFK                 SFO   8437
 10925            OGG                 HNL   8313
 9536             HNL                 OGG   8282
 10149            LAX                 ORD   8256
 7887             ATL                 LGA   8234
 10204            LGA                 ATL   8215
 7892             ATL                 MCO   8202
 10372            MCO                 ATL   8202
 11866            SFO                 LAS   7995
 11082            ORD                 LAX   7941
 10083            LAS                 SFO   7870
 8989             DFW                 ORD   7870
 11797            SEA                 LAX   7765
 10164            LAX                 SEA   7732
 8168             BOS                 DCA   7687
 8693             DCA                 BOS   7686
 11044            ORD                 DFW   7677
 7854             ATL                 FLL   7419
 9337             FLL                 ATL   7411
 11342            PHX                 LAX   7380
 11887            SFO                 ORD   7380
 11138            ORD                 SFO   7378
 10153            LAX                 PHX   7330
 8851             DEN                 PHX   7211
 11322            PHX                 DEN   7204
 10208            LGA                 BOS   7100
 8182             BOS                 LGA   7096
 12220            TPA                 ATL   7083
 7955             ATL                 TPA   7076
 8891             DFW                 ATL   7060
 7839             ATL                 DFW   7056, :Dataset   [Airport])

type(airports_dataset)

holoviews.core.data.Dataset

Applying Colors to the Chord Diagram

• https://docs.bokeh.org/en/latest/docs/reference/palettes.html

node_cmap은 노드에 적용할 팔레트를 나타내며, edge_color는 코드 다이어그램 엣지에 적용할 팔레트를 나타냅니다.

%%opts Chord [height=500 width=500 title="Flights between airports" labels="Airport"]
%%opts Chord (node_color="Airport" node_cmap="Category20" edge_color="ORIGIN_AIRPORT" edge_cmap='Category20')
chord = hv.Chord((df_between_airports, airports_dataset))
chord

다른 팔레트 적용

%%opts Chord [height=500 width=500 title="Flights between airports" labels="Airport"]
%%opts Chord (node_color="Airport" node_cmap="Category20" edge_color="ORIGIN_AIRPORT" edge_cmap='Bokeh')
chord = hv.Chord((df_between_airports, airports_dataset))
chord

Confirming the Relationships in the Chord Diagram

이 글의 마지막 섹션에서는 코드 다이어그램이 올바른 정보를 표시하는지 확인하고 싶습니다. JFK를 선택하면, 항공편이 LAX와 SFO로 날아가고 있다는 것을 알 수 있습니다.

df_between_airports.query('ORIGIN_AIRPORT == "JFK"')

	ORIGIN_AIRPORT	DESTINATION_AIRPORT	COUNT
9949	JFK	LAX	12016
9974	JFK	SFO	8437

df_between_airports.query('ORIGIN_AIRPORT == "ORD"')

	ORIGIN_AIRPORT	DESTINATION_AIRPORT	COUNT
11085	ORD	LGA	9575
11082	ORD	LAX	7941
11044	ORD	DFW	7677
11138	ORD	SFO	7378